Model Struct Reference

Scene model used to calibrate Simd::Motion::Detector. More...

#include <SimdMotion.hpp>

Public Member Functions
	Model (const Model &model)
	Model (const FSize &size_=FSize(0.1, 0.1), const FPoints &roi_=FPoints())
	Model (const FSize &size_, const View &mask_)
Model &	operator= (const Model &model)

Data Fields
FSize	size
	Minimum object size in ONVIF units [0, 2]. Default (0.1, 0.1). Shooting-star example uses (0.01, 0.01).
FPoints	roi
	ROI polygon in ONVIF coordinates [-1, 1]. Used when mask is not Gray8. Empty (fewer than 3 points) means the full screen.
View	mask
	ROI mask. If format is View::Gray8, Detector uses it instead of roi (non-zero pixels are inside).

Static Public Attributes
static const uint8_t	EMPTY = 0
	Mask value outside ROI. Detector fills the polygon mask with this value.
static const uint8_t	ROI = 255
	Mask value inside ROI. Typical usage writes ROI (or any non-zero) into Model::mask.

Detailed Description

Scene model used to calibrate Simd::Motion::Detector.

Typical usage constructs Model (or uses the default), optionally sets size / roi / mask, and passes it to Detector::SetModel before NextFrame.

size is the minimum object size in ONVIF units. Default (0.1, 0.1) is about 0.25% of the screen area. The shooting-star example uses (0.01, 0.01).

ROI is either a polygon (roi, at least 3 ONVIF vertices) or a Gray8 mask. If mask.format is View::Gray8, Detector resizes it to the frame and treats non-zero pixels as inside ROI (EMPTY = 0, ROI = 255). Otherwise the polygon is used; fewer than 3 vertices mean the full screen.

Constructor & Destructor Documentation

Model() [1/3]

Model

(

const Model &

model

)

Copy constructor of Model.

Copies size, roi and, if it is Gray8, a deep copy of mask.

Parameters

[in]

model

- other model.

Model() [2/3]

Model	(	const FSize &	size_ = FSize(0.1, 0.1),
		const FPoints &	roi_ = FPoints()
	)

Constructs Model from a minimum object size and an ROI polygon.

If roi_ has fewer than 3 points, roi is set to the full-screen rectangle (-1, 1), (1, 1), (1, -1), (-1, -1). mask is left empty, so Detector uses the polygon.

Parameters

[in]	size_	- a minimum object size in ONVIF units. Default is (0.1, 0.1) ~ 0.25% of screen area.
[in]	roi_	- a ROI polygon in ONVIF coordinates. Empty by default (full screen).

Model() [3/3]

Model	(	const FSize &	size_,
		const View &	mask_
	)

Constructs Model from a minimum object size and an ROI mask.

If mask_ is Gray8, it is deep-copied to mask. Otherwise roi is set to the full screen. Typical usage creates a small Gray8 mask (for example 20x20, left half filled with Model::ROI) and calls SetModel.

Parameters

[in]	size_	- a minimum object size in ONVIF units.
[in]	mask_	- a ROI mask. It must be 8-bit gray image (View::Gray8).

Member Function Documentation

operator=()

Model & operator=

(

const Model &

model

)

Copy assignment of Model.

Copies size, roi and, if it is Gray8, a deep copy of mask.

Parameters

[in]

model

- other model.